pmttyji

• llama: limit max outputs of `llama_context` by am17an · Pull Request #23861 · ggml-org/llama.cpp

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 44 comments

• Mellum & Granite Embedding models are ready on llama.cpp

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 1 comments

• Open Models - May 2026

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 27 comments

• Model: Support Step3.7-Flash by forforever73 · Pull Request #23845 · ggml-org/llama.cpp

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 6 comments

• nvidia/Qwen3.6-35B-A3B-NVFP4 · Hugging Face

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 51 comments

• OpenMOSS-Team/MOSS-TTS-v1.5 · Hugging Face

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 19 comments

• We need some polls on many topics - 2026

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 13 comments

• model : add support for talkie-1930-13b by niklassheth · Pull Request #22596 · ggml-org/llama.cpp

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 7 comments

• OSCAR RotationZoo - Offline Spectral Covariance-Aware Rotation for 2-bit KV Cache Quantization

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 8 comments

• Full Attention Strikes Back: Transferring Full Attention into Sparse within Hundred Training Steps

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 27 comments

• Is Granite-4.1-30b Overshadowed by Qwen3.6 & Gemma4 models?

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 69 comments

• Next year we're getting 0.5T model from Grok

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 200 comments

• CUDA: add fast walsh-hadamard transform by am17an · Pull Request #23615 · ggml-org/llama.cpp

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 16 comments

• Jackrong/Qwopus3.5-9B-Coder-GGUF · Hugging Face

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 72 comments

• Add MiniCPM5 tokenizer support by zhangtao2-1 · Pull Request #23384 · ggml-org/llama.cpp

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 0 comments

• Compilation of recent findings which could save some memory on increase performance

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 14 comments

• meituan-longcat/LongCat-Video-Avatar-1.5 · Hugging Face

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 15 comments

• TurboQuant isn’t just for KV: Qwen3.5-27B at near-Q4_0 quality, about 10% smaller, and finally fitting on my 16GB 5060 Ti

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 161 comments

• numind/NuExtract3 · Hugging Face

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 6 comments

• ByteDance-Seed/Cola-DLM · Hugging Face

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 10 comments

• Open Models - April 2026 - One of the best months of all time for Local LLMs?

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 153 comments

• LlamaStation v0.9 — llama.cpp GUI for Windows with multi-backend support, TurboQuant, MTP and more

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 21 comments

• Sarvam-30b-quantized - Need 1-bit version GGUF

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 26 comments

• Finalizing my New Desktop Rig 96GB VRAM + 128GB RAM - 3rd GPU possible in this setup?

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 47 comments

• LlamaStation v0.9 — llama.cpp GUI for Windows with multi-backend support, TurboQuant, MTP and more

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 0 comments

• CohereLabs/command-a-plus-05-2026-bf16 · Hugging Face

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 6 comments

• Finalizing my New Desktop Rig 144 GB VRAM + 128GB RAM - 3rd GPU possible in this setup?

  Posted by pmttyji@reddit | buildapc | View on Reddit | 9 comments

• Thinking of getting two NVIDIA RTX Pro 4000 Blackwell (2x24 = 48GB), Any cons?

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 92 comments

• How many of you tried BeeLlama.cpp? How's it? Agentic coding possible with 8GB VRAM?

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 65 comments

• Why some Github projects only support wrappers instead of llama.cpp?

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 34 comments

• internlm/Intern-S2-Preview · Hugging Face

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 11 comments

• MagenticLite is here: A full-stack agentic experience powered by Small Models - Fara-1.5 4B, 9B & 27B

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 6 comments

• sensenova/SenseNova-U1-A3B-MoT · Hugging Face

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 5 comments

• New models possibly from Baidu (ERNIE) this month?

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 14 comments

• AIDC-AI/Ovis2.6-80B-A3B · Hugging Face

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 28 comments

• A Dark-Money Campaign Is Paying Influencers to Frame Chinese AI as a Threat

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 171 comments

• llama.cpp - Custom Optimized Builds?

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 20 comments

• Poor GPU Club : Tried Bonsai-8B on CPU & CUDA

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 36 comments

• inclusionAI/Ling-2.6-1T · Hugging Face

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 16 comments

• Peanut - Text to Image Model (Open Weights coming soon)

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 27 comments

• Experts-Volunteers needed for Vulkan on ik_llama.cpp

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 13 comments

• GitHub - warpdotdev/warp: Warp is an agentic development environment, born out of the terminal.

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 2 comments

• Something from Mistral (Vibe) tomorrow

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 92 comments

• Remote agents in Vibe. Powered by Mistral Medium 3.5.

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 0 comments

• SenseNova-U1: Unifying Multimodal Understanding and Generation with NEO-Unify Architecture

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 2 comments

• Recent Open models from last 6 Months - Nov 2025 - Apr 2026

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 45 comments

• Ternary Bonsai: Top intelligence at 1.58 bits

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 89 comments

• ggml-cpu: Optimized x86 and generic cpu q1_0 dot (follow up) by pl752 · Pull Request #21636 · ggml-org/llama.cpp

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 3 comments

• Excerpts from *The Technological Republic*

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 0 comments

• Mixture-of-Depths Attention - arXiv

  Posted by pmttyji@reddit | LocalLLaMA | View on Reddit | 1 comments